Hindawi

Mathematical Problems in Engineering

Volume 2023, Article ID 2889435, 9 pages
https://doi.org/10.1155/2023/2889435

Review Article
The Dynamic Model and Control Algorithm for the Active
Suspension System

Duc Ngoc Nguyen and Tuan Anh Nguyen

Faculty of the Mechanical Engineering, Tuyloi University, Hanoi, Vietnam

Correspondence should be addressed to Duc Ngoc Nguyen; [email protected]

Received 18 August 2022; Revised 7 January 2023; Accepted 25 January 2023; Published 7 February 2023

Academic Editor: Xingling Shao

Copyright © 2023 Duc Ngoc Nguyen and Tuan Anh Nguyen. Tis is an open access article distributed under the Creative
Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the
original work is properly cited.
Road surface roughness is the leading cause of vehicle oscillation. Te suspension system is used to dampen these oscillations. Te
active suspension system equipped with a hydraulic actuator is more efcient than the passive one. Terefore, it is used to replace
the passive suspension system. Te article reviews and analyses models and control algorithms for active suspension systems. In
this article, the author mentioned three dynamic models commonly used to simulate vehicle oscillations: a quarter-dynamic
model, a half-dynamic model, and a fully dynamic model. Each hydraulic actuator can be considered a state variable in the
dynamic model. Besides, the control algorithm for the suspension system is signifcant. Algorithms like PID (proportio-
nal–integral–derivative) and LQR (linear quadratic regulator) will ft the linear model. In contrast, the nonlinear model’s al-
gorithms, such as SMC (sliding mode control), fuzzy, and ANN (artifcial neural network), will perform better. Overall, vehicle
oscillation can be signifcantly improved once an active suspension system is used. Te contents analysed in this article will be the
database for selecting control models and algorithms by other researchers in the future.

1. Introduction Researchers have recently proposed using electronic

suspension systems to replace mechanical suspension sys-
Te suspension system is a critical component of the vehicle. tems. Electronic suspension systems have a more complex
Te suspension system controls the stability and comfort of structure, and at the same time, their cost is also higher.
the vehicle. Te suspension system is not only equipped for However, the smoothness efect it brings is also more sig-
cars but is also used for all vehicles, such as trains, mo- nifcant. Tere are three commonly mentioned electronic
torcycles, airplanes, etc. A basic car suspension system suspension systems: the air suspension system, the semi-
structure usually has three main components: coil springs, active suspension system, and the active suspension system.
dampers, and a lever arm [1]. Te number of lever arms will Air suspension is a rather complicated system. Te air
depend on the specifc type of suspension. For the springs of the suspension system can be changed in stifness
McPherson suspension, only the lower lever arm is used by changing the pressure inside the chambers of the air
(Figure 1). Some models can use multilink bars to replace the balloon [4]. According to Yin et al., this suspension system
lever arm. For truck models, the coil spring is replaced by a often uses solenoid valves (ON-OFF valves) to regulate the
leaf spring [2]. Te above suspension system is called the air supply [5]. Te cost of the air suspension system is usually
passive (mechanical) suspension system [3]. Te structure of relatively high. Terefore, it is often used for high-end cars.
a mechanical suspension is usually quite simple. However, Another suspension system that can change the stifness of
its working efciency is not high. Vehicle body oscillations the dampers is called a semi-active suspension system. Tis
are controlled in a completely passive way. As a result, suspension system uses electromagnetic dampers to replace
stability and comfort can be lost if vehicle body oscillations conventional linear dampers [6]. According to Yang et al.,
are too great. the current provided to the damper can change the
2 Mathematical Problems in Engineering

arrangement of the surrounding metal particles. Terefore,

Damper
the viscosity of the liquid inside will also change [7]. A semi-
Coil spring
active suspension system has a more straightforward
structure than an air suspension system. Terefore, its cost is
also lower. However, its oscillatory conditioning efciency is
also lower. Another solution has also been proposed, which
is to use an active suspension system. According to Huang
et al., the active suspension system has an additional hy- Hub component
draulic actuator [8]. Te hydraulic actuator operates based
on the displacement of the servo-valve [9]. Te pump
pressurizes the fuid before it is directed into the hydraulic
cylinder. Te idea of using linear pumps for hydraulic ac-
tuators was proposed by Lee et al. [10]. Te active suspension
system has a more complex structure than the semi-active
one. However, the hydraulic actuator can generate force to
act on the vehicle body, which helps reduce vehicle oscil-
lation. Terefore, its performance will be higher. In addition,
the active suspension system also helps to limit the rollover
phenomenon. Tis was pointed out in [11] by Guan et al.,
Lower lever arm
but it cannot replace the function of the stabilizer bar [12].
Te active suspension system is commonly used in high-end Figure 1: McPherson suspension system.
cars. In particular, electric vehicles using motor-in-wheel are
also equipped with an active suspension system [13]. algorithms. In the conclusion part, several conclusions will
Tanks to the suspension system, the car’s oscillation can be provided.
be improved. Many criteria have been used to evaluate the
smoothness of the vehicle when moving. In [14], Lan and Xu
used the criteria of vertical displacement and vertical ac- 2. Model and Control
celeration of the sprung mass to evaluate the stability and
2.1. Vehicle Dynamic Models. Dynamic models can explain
comfort once the vehicle oscillates. Many other studies also
automobile oscillations. Tese models represent the entire
used these values [15–17]. Te maximum and average values
vehicle structure’s simplicity. Tere are three typical models:
will be considered for comparison between oscillation sit-
the quarter-dynamic model, the half-dynamic model, and
uations. Besides, the average value can be calculated
the fully dynamic model.
according to the RMS (root mean square) criterion [18].
Some other studies also use DLC (dynamic load coefcient)
criteria to evaluate the interaction between the wheel and the 2.1.1. Te Quarter-Dynamic Model. A quarter-dynamic
road surface when the vehicle body oscillates [19]. When model is commonly used in studies of oscillation control for
using a half-dynamic or spatial dynamic model, some values, suspension systems. According to Wang et al., this model
such as roll angle, pitch angle, etc., should also be considered includes only two masses: the sprung mass, ms, and the
[20]. Te output values of the simulation problem can be unsprung mass, mu [24]. Te suspension system is modelled
reviewed in the time domain [21, 22] or the frequency as a spring and a shock absorber. Te tires are also modelled
domain [23]. similarly to the suspension system. For vehicles with an
An active suspension system improves vehicle ride active suspension system, an actuator is located between the
comfort by generating impact force that is transmitted to the sprung mass and the unsprung mass (Figure 2). Te
sprung and unsprung masses. If the value of this force is equations describing the oscillations of a quarter-dynamic
small, the system’s response is not good. Tis means that the model are shown as follows:
car’s vibration has yet to improve much. If the impact force
is large, the ride comfort can be further improved. However, ms €zs � K zu − zs 􏼁 + C z_u − z_s 􏼁 + Fa ,
(1)
a more signifcant impact force will cause a change in the mu €zu � KT zr − zu 􏼁 − K zu − zs 􏼁 − C z_u − z_s 􏼁 − Fa ,
dynamic load at the wheel. Once the value of the dynamic
force at the wheel is reduced to zero, the wheel may be lifted where K is the stifness of the spring, KT is the stifness of the
of the road, and instability will occur. It is difcult to satisfy tire, and C is the stifness of the damper.
both conditions of smoothness and stability when studying Overall, this model is quite simple. Terefore, it is
control algorithms for active suspension. perfectly suitable for simulating control algorithms for
Te main content of this article is to review the models suspension systems. Tis model is also used in [25] by Chen
and control algorithms for the active suspension system. Te et al. However, the authors hypothesized that the controller
article’s structure consists of three parts: introduction, would generate a force to act directly on both parts of the
control and model, and conclusion. In the frst part, active mass. Te infuence of hydraulic actuators has been
suspension system ideas are introduced. Te second part neglected [25, 26]. Tis simplifes the control problem. Te
presents a review and analysis of control models and problem’s accuracy is enhanced once the hydraulic actuator
Mathematical Problems in Engineering 3

zs
zs
ms
Car body
xv

Active Fs Fd Pr
u
suspension P2
Q2
zu
Ps
Unsprung mass mu Q1

Pr P1 zu
Ft Fb
Tire spool
zr
Servo valve Hydraulic Cylinder
Figure 3: Hydraulic actuator [27].
Figure 2: A quarter-dynamic model [24].
Te dynamic model of the actuator is quite complex. Tis
is considered in the control model (Figure 3). According to model should be linearized for some control algorithms
Kilicaslan, the operating principle of hydraulic actuators can using higher-order derivatives. Te process of linearizing
be described through complex nonlinear equations [27]. hydraulic actuators was performed by Nguyen et al. [28]. Te
Tese equations are presented as follows: equation describing the operating principle of the actuator is
given in a linear form as (3). Te linear diferential equation
Fa � APL ,
of the actuator is also used in many other studies [29, 30].
PL � P 1 − P 2 , F_ a � c1 i(t) − c2 Fa − c3 z_s − z_u 􏼁, (3)

Vt _ where ci is the coefcient.

P � QL − Ct PL − A z_s − z_u 􏼁, (2)
4βe L
􏽳����������������
1 2.1.2. Te Half-Dynamic Model. A half-dynamic model of a
QL � Cd xv w 􏼂Ps − sgn xv 􏼁PL 􏼃, vehicle has four degrees of freedom (Figure 4). It can be roll
ρ model or pitch model. Tis model considers the infuence of
the vehicle’s body roll or pitch angles.
where Vt: efective volume of the hydraulic actuator, βe: bulk
Considering the pitch model, the equations describing
modulus of the fuid, Ct: leakage constant, QL: hydraulic load
the oscillations of the vehicle are established based on
fow, Cd: discharge coefcient, w: servo valve area gradient,
Newton’s second law [31].
xv : servo valve spool displacement, and ρ: fuid density.

ms €
zc + ksf 􏼐zsf − zuf 􏼑 + csf 􏼐z_sf − z_uf 􏼑 + ksr zsr − zur 􏼁 + csr z_sr − z_ur 􏼁 � uf + ur ,
€ − l1 ksf 􏼐zsf − zuf 􏼑 − l1 csf 􏼐z_sf − z_uf 􏼑 + l2 ksr zsr − zur 􏼁 + l2 csr z_sr − z_ur 􏼁 � −l1 uf + l2 ur ,
Iφ φ
(4)
muf €
zuf − ksf 􏼐zsf − zuf 􏼑 − csf 􏼐z_sf − z_uf 􏼑 + ktf 􏼐zuf − zrf 􏼑 � −uf ,
mur €
zur − ksr zsr − zur 􏼁 − csr z_sr − z_ur 􏼁 + ktr zur − zrr 􏼁 � −ur ,

where ksf is the stifness of the front spring, ksr is the stifness of 2.1.3. Te Fully Dynamic Model. Another model used to
the rear spring, ktf is the stifness of the front tire, ktr is the describe vehicle oscillations is the fully dynamic model
stifness of the rear tire, csf is the stifness of the front damper, csr (spatial dynamic model). Tis model consists of 5 masses
is the stifness of the rear damper, uf is the control force of the and 7 degrees of freedom (Figure 5). Overall, this model is
front actuator, and ur is the control force of the rear actuator. quite complex. Te controllers of each actuator may have
Tis model uses two actuators at the vehicle’s front and signal conficts when working. However, this model pro-
rear. Two independent controllers can control the two ac- vides all the necessary elements to evaluate the vehicle’s
tuators. In some cases, these two controllers can also interact oscillation.
with each other. Tis is a complete model, but not too Te system of equations describing the oscillations of the
complicated. It is also commonly used in many vehicle spatial model was established by Sun et al. in [32].
oscillation simulation problems.
4 Mathematical Problems in Engineering

v
l1 Zc l2
z
y
θ Zs1
φ Zs3 φ
ms, Iφ Zs4 Zb x Zs2
c
d
b a

Zsf Zsr
Fa1
uf (t – d (t)) uf (t) ur (t – d (t)) Fa4 Fa3
Fa2 c1 k1
ksf Delay Controller ksr Delay Controller c3 k3 Zu3 Zu1
Zu4 c4 k4 mu1
csf csr mu3 c2 k2
mu4 Zu2
Zuf Zur mu2
muf mur kt3 kt1
kt4 Zr3
Zr4 kt2 Zr1
Zr2
ktf ktr
Zrf Zrr

Figure 5: Te fully dynamic model [32].

Figure 4: Half-dynamic model [31].

ms €
zs + 􏽘 Fsi + Fdi − Fai + Fati1 + Fati2 + Fati3 + Fati4 􏼁 � 0,
Iy €θ + a Fs1 + Fd1 + Fs2 + Fd2 + Fati1 + Fati2 􏼁 − b Fs3 + Fd3 + Fs4 + Fd4 + Fati3 + Fati4 􏼁 − Mθ � 0,
€ − d Fs2 + Fd2 + Fati2 + Fs4 + Fd4 + Fati4 􏼁 + c Fs1 + Fd1 + Fati1 + Fs3 + Fd3 + Fati3 􏼁 − Mφ � 0,
Ix φ
mu1 €
zu1 − Fs1 − Fd1 + Ft1 − Fa1 � 0, (5)
mu2 €
zu2 − Fs2 − Fd2 + Ft2 − Fa2 � 0,
mu3 €
zu3 − Fs3 − Fd3 + Ft3 − Fa3 � 0,
mu4 €
zu4 − Fs4 − Fd4 + Ft4 − Fa4 � 0,

where Fs: spring force, Fd: damper force, Ft: tire force, Fa: called the LQG (Linear Quadratic Gaussian) algorithm [43].
control force, Mφ: roll moment, and Mθ: Pitch moment. Te parameters of the control matrix can be selected em-
Te binding forces of the dynamic model are calculated pirically or by intelligent algorithms. Tese parameters are
more specifcally in [32]. fxed. Te LPV control algorithm can be used if variable
parameters are used instead [44].
2.2. Control Algorithms. Many articles on control algorithms 1 T T
J � lim 􏽚 􏼐X QX + UT RU + 2XT NU􏼑dt ⟶ Min.
for suspension systems have been published. Te choice of T⟶∞ T 0
algorithms will depend on the views of the researchers. If the (6)
pavement excitations are linear, the oscillating system is
considered to be a linear system. Ten, algorithms like PID, A result of using the PID algorithm to control the active
LPV (linear parameter varying), LQR, etc., can be used. In suspension is shown in Figure 6. Tis is the change in vehicle
[33], Liu et al. proposed using the PID algorithm for a body acceleration over simulation time. Te acceleration
quarter-dynamic suspension system model. If the difer- value is maximum if the vehicle uses only a passive sus-
ential stage is omitted, this algorithm has only two com- pension system. Tis value can be reduced once the PID
ponents, KP and KI [34]. If the integral is ideal, the PID algorithm is used. In [33], Liu et al. used the BPNN (back
algorithm becomes PD [35]. Te parameters of the PID propagation neural network) algorithm to determine the
algorithm can be self-tuning or tuned by intelligent algo- PID controller’s parameters. Terefore, the performance of
rithms such as fuzzy, PSO (particle swam optimization), GA the controller can be better improved. In [39], Nguyen
(genetic algorithm), etc. [36–38]. Tis is only a SISO (single compared the use of a single PID controller and a double
input–single output) algorithm, i.e., it can only control one PID controller. According to the results shown in Figure 7,
object. If the number of objects to be controlled is greater, the displacement of the vehicle body is signifcantly reduced
more controllers and actuators should be used [39]. How- once the double PID controller is used. However, improper
ever, this can cause internal efects on the system. Terefore, selection of parameters can lead to the oscillation phase
the better solution is to use the LQR algorithm for the multi- diference.
object system. In [40], Nguyen et al. presented the LQR Te pavement stimuli are random; therefore, the vehi-
control algorithm for the MIMO (multi-input–multioutput) cle’s oscillation is nonlinear. For nonlinear systems, more
system with the model of fve state variables. Tis algorithm complex algorithms should be used. In [45], Bai and Wang
needs to use matrices describing the oscillation state of the proposed using optimally robust algorithms for suspensions
dynamic model. Minimizing the cost function, like in with uncertainties. Tis algorithm was established based on
equation (6), will make the system more stable [41]. A the combination of the LQR algorithm and the SMC al-
Gaussian flter is integrated to remove noise [42]. Ten, it is gorithm. Besides, Wei et al. also showed a new method for
Mathematical Problems in Engineering 5

300
3
2 200
Body acceleration (m/s2)

1 100

u (t)
0
–1
–100
–2
–200
–3
–4 –300
0 5 10 15 20 0 1 2 3 4 5 6 7 8
Time (s)
Time (s)
Passive Figure 8: Force of the actuator [53].
PID
BP-PID
algorithm and the PID algorithm to become a synthesis
Figure 6: Body acceleration [33]. algorithm [54]. Te SMC algorithm’s output signal will be
the PID algorithm’s input signal. Terefore, the phenome-
non of “chattering” has been signifcantly reduced. Another
50 way to limit the phenomenon of “chattering” has also been
Displacement of the sprung mass (mm)

40 described by Wang et al., as shown in [55]. Te efciency of

30 the SMC algorithm is very high, which is proved by Nguyen
20 in [54]. Displacement of the vehicle body can be up to 61.96
10
(mm) if the car only uses conventional mechanical sus-
pension (Figure 9). When the PID controller controls the
0
active suspension, this value is reduced to 23.35 (mm). Tis
-10
value decreases to only 7.73 (mm) once the SMC algorithm is
-20 applied. Furthermore, if the PID algorithm and the SMC
-30 algorithm are combined, the maximum displacement value
-40 of the sprung mass will not exceed 1.00 (mm). Tis is an ideal
-50 result. However, the design of this algorithm is also quite
0 5 10 15 20 25 30
complicated. Sliding control algorithms can be applied to
Time (s)
many system models, such as leg robots [56] or quadrotor
Active DC UAVs [57]. Te performance that this algorithm brings is
Active SC outstanding. For uncertain systems, it is necessary to use
Passive
composite controllers. V. Ghafari and S. Mobayen et al.
Figure 7: Displacement of the sprung mass [39]. propose a robust tracking nonlinear feedback controller for
the system [58], which results in high efciency.
Several intelligent control algorithms have also been
the H∞ controller [46]. According to Wei et al., this con- used to control the suspension. In [59], Ozbek et al. used the
troller is established based on solving linear and related fuzzy algorithm to control the spatial model. A fuzzy al-
matrices. Besides, using adaptive control algorithms for gorithm allows for defning an intermediate state when the
harmonic oscillations also brings high efciency [47]. In system is working. Tis algorithm is determined based on
[48], Gang proposed using a fully model’s SMC algorithm the membership function that has been designed before [60].
for suspension. Tis algorithm can also be used for half- or Diferent researchers’ points of view formulate fuzzy rules.
quarter-models [49, 50]. Te SMC algorithm needs to use Te fuzzy algorithm can also be combined with the SMC
the higher-order derivative of the output signals, so the use algorithm to eliminate the “chattering” phenomenon, as
of a linear model of the actuator is necessary [29]. For this shown by Shaer et al. [61]. Several algorithms using ANN
algorithm, the sliding surface is an essential component, so it have also been proposed to control the suspension system. In
needs to be set up carefully [51]. Te input signal from the [62], Zhang et al. used the AFTNN (adaptive fnite-time
controller can be received by sensors or cameras [52]. A neural network) control algorithm for the suspension model
drawback still exists when using the SMC algorithm, which with three degrees of freedom. More recently, an adaptive
is the phenomenon of “chattering.” Tis phenomenon will control algorithm using a saturated neural network has also
cause the signal to vibrate continuously [53]. Terefore, it been proposed by Wang and Li [63]. In addition, many PSO
can afect the quality of the system. A simulation result of algorithms that use animal behaviour are also used for
this phenomenon is shown in Figure 8. Accordingly, the suspension systems [64–66]. Tese algorithms also bring
control signal oscillates continuously at a high frequency high efciency to the system. Moreover, some control al-
throughout the simulation time. Tis is not good for the gorithms for mobile robots and quadrotors should also be
system. To solve this problem, Nguyen combined the SMC considered to control the active suspension [67–70].
6 Mathematical Problems in Engineering

60

Displacement of the sprung mass (mm)

40

20

-20

-40

-60
0 5 10 15
Time (s)

SMC-PID PID
SMC Passive

Figure 9: Displacement of the sprung mass [54].

Table 1: Control methods for the suspension system.

Methods Objects of application Characteristics
Applicable to linear systems with Te design process is simple, low-cost, and widely used in
PID
only a single input and output. industry. It is necessary to choose the parameters appropriately.
Te design process is simple; however, it is necessary to use a
Applicable to linear systems with
LQR/LQG matrix of oscillations instead of the usual system of diferential
multiple inputs and outputs.
equations.
High stability and good performance. It is necessary to use the
Applicable to complex nonlinear
SMC signal of the higher-order derivative while the “chattering”
systems.
phenomenon can occur. Te design process is very complex.
Applicable to complex nonlinear Good responsiveness to the continuous change of input
Robust, adaptive
systems. parameters. Te design process is complicated.
Well-adapted to the continuous change of input parameters.
Applicable to multiple complex
Fuzzy, ANN, PSO, etc. Te design of algorithms is challenging and depends a lot on the
systems.
designer’s point of view.
Combination methods (SMC-PID, High stability and good adaptability. Te combination process
Applicable to multiple complex
AFTNN, Fuzzy-SMC, Fuzzy-PID, is very difcult, and there may be conficts between the
systems.
etc.) component algorithms.

Table 1 shows a comparison of the diferent ways to author’s point of view, three dynamic models are used to
control an active suspension system. simulate vehicle oscillations: the quarter, the half, and the
spatial. Among them, the quarter model is quite commonly
used. Te hydraulic actuator is also considered a state
3. Conclusions
variable of the dynamic model. Control algorithms such as
Te suspension system plays an essential role in ensuring the PID, LQR, and LPV may apply if the system oscillation is
stability and comfort of the vehicle. Te structure of the passive considered linear. Tese algorithms are quite simple.
suspension system has only three basic components, so it However, the response is good only in a few specifc cases. In
cannot meet the requirements for smoothness in motion. contrast, more complex algorithms such as robust control,
Terefore, an active suspension system replaces a conventional adaptive control, SMC, Fuzzy, ANN, etc., can be suitable for
passive suspension system. Te hydraulic actuator is a com- nonlinear systems at all times. Te choice of algorithms is
ponent that may be found in the active suspension system. Te entirely up to each researcher’s point of view.
operation of this actuator is determined by whether or not the Te input of the simulation problem is the stimuli from
servo valves are open or closed. It can exert force on both the the road surface; the output usually includes the values of
sprung and unsprung mass of the vehicle, which helps to lessen displacement, acceleration, and roll angle of the vehicle
oscillation. However, the structure of the active suspension body. Tese values can be evaluated according to RMS, DLC,
system will be more complicated than that of the passive maximum value, etc. Overall, when active suspension is
suspension system. used, ride comfort and stability can improve. Te contri-
Tis article introduces and reviews models and control bution of this article can make it easier for researchers to
algorithms for the active suspension system. From the choose more suitable control models and algorithms. Today,
Mathematical Problems in Engineering 7

several challenges related to controlling issues for the active stifness component,” Mechanical Systems and Signal Pro-
suspension system still exist. For example, if the ride comfort cessing, vol. 147, Article ID 107071, 2021.
is improved, the interaction between the wheels and the road [8] Y. Huang, J. Na, X. Wu, and G. Gao, “Approximation-free
surface may be negatively afected. Conversely, if the wheels control for vehicle active suspensions with hydraulic actua-
are controlled to increase road holding, smoothness and tor,” IEEE Transactions on Industrial Electronics, vol. 65, no. 9,
pp. 7258–7267, 2018.
comfort can be lost. Both of the above factors are very
[9] L. Xiao, M. Wang, B. Zhang, and Z. Zhong, “Vehicle roll
important for the suspension system, and it is not easy to stability control with active roll-resistant electro-hydraulic
satisfy them simultaneously. suspension,” Frontiers of Mechanical Engineering, vol. 15,
no. 1, pp. 43–54, 2020.
4. Future Recommendation [10] J. Lee, K. Oh, and K. Yi, “A novel approach to design and
control of an active suspension using linear pump con-
In the future, the suspension system on cars will be further trol–based hydraulic system,” Proceedings of the Institution of
developed to be more robust. Te suspension system control Mechanical Engineers - Part D: Journal of Automobile Engi-
algorithms need to satisfy ride comfort and road-holding neering, vol. 234, no. 5, pp. 1224–1248, 2020.
requirements. To do this, combining many complex non- [11] Q. Guan, A. Gong, M. Hu, Z. Liao, and X. Chen, “Anti-
rollover warning control of dump truck lifting operation
linear algorithms or using intelligent algorithms is necessary.
based on active suspension,” Journal of Control, Automation
Besides, it is necessary to carry out experiments to evaluate and Electrical Systems, vol. 32, no. 1, pp. 109–119, 2021.
the actual efectiveness of the controllers for the car sus- [12] D. N. Nguyen, T. A. Nguyen, and N. D. Dang, “A complex
pension system. Another solution is to use many inde- rollover dynamics model with active stabilizer bar controlled
pendent controllers for many diferent purposes. by the fuzzy algorithm,” Heliyon, vol. 8, no. 11, Article ID
e11715, 2022.
Data Availability [13] X. Lin, J. Wang, and J. Yang, “Development of robust
guaranteed cost mixed control system for active suspension of
Te data used to support the fndings of this study are in- in-wheel-drive electric vehicles,” Mathematical Problems in
cluded within the article. Engineering, vol. 2022, Article ID 4628539, 16 pages, 2022.
[14] J. Lan and T. Xu, “Adaptive fnite-time fault-tolerant control
for half-vehicle active suspension systems with output con-
Conflicts of Interest straints and random actuator failures,” Complexity, vol. 2021,
Article ID 5964034, 16 pages, 2021.
Te authors declare that there are no conficts of interest [15] S. J. Shao, D. Jing, and C. B. Ren, “Multiobjective optimization
regarding the publication of this article. of nonlinear active suspension system with time-delayed
feedback,” Mathematical Problems in Engineering, vol. 2020,
Article ID 9526359, 11 pages, 2020.
References [16] L. X. Guo and L. P. Zhang, “Constrained active suspension
[1] T. E. Putra, Husaini, and M. Ikbal, “Automotive suspension control with parameter uncertainty for non-stationary run-
component behaviors driven on fat and rough road surfaces,” ning based on LMI optimization,” Journal of Vibration En-
Heliyon, vol. 7, Article ID e07528, 2021. gineering & Technologies, vol. 6, pp. 441–451, 2018.
[2] M. Stanco and M. Kowalczyk, “Analysis of experimental [17] D. N. Nguyen and T. A. Nguyen, “Enhancing the performance
results regarding the selection of spring elements in the front of the vehicle active suspension system by an Optimal Sliding
suspension of a four-axle truck,” Materials, vol. 15, no. 4, Mode Control algorithm,” PLoS One, vol. 17, no. 12, Article ID
p. 1539, 2022. e0278387, 2022.
[3] M. Mohsen, H. Kamel, A. M. Sharaf, and S. M. E. Demerdash, [18] F. R. Kou, D. D. Wei, and L. Tian, “Multimode coordination
“Optimal design of passive suspension system of a 6 × 6 multi- control of a hybrid active suspension,” Shock and Vibration,
wheeled all-terrain vehicle using genetic algorithm,” Inter- vol. 2018, Article ID 6378023, 16 pages, 2018.
national Journal of Heavy Vehicle Systems, vol. 25, no. 3/4, [19] Y. Chen, J. He, M. King, W. Chen, and W. Zhang, “Efect of
pp. 508–533, 2018. driving conditions and suspension parameters on dynamic
[4] T. A. Nguyen, “Advance the stability of the vehicle by using load-sharing of longitudinal-connected air suspensions,”
the pneumatic suspension system integrated with the hy- Science China Technological Sciences, vol. 56, no. 3, pp. 666–
draulic actuator,” Latin American Journal of Solids and 676, 2013.
Structures, vol. 18, no. 7, Article ID e403, 2021. [20] Y. J. Liu, Y. Zhang, L. Liu, S. Tong, and C. L. P. Chen,
[5] C. Yin, X. Zhai, X. Sun, S. Wang, and P. K. Wong, “Design and “Adaptive fnite-time control for half-vehicle active suspen-
performance research of a hydro-pneumatic suspension with sion systems with uncertain dynamics,” IEEE/ASME Trans-
variable damping and stifness characteristics,” Journal of actions on Mechatronics, vol. 26, no. 1, pp. 1–178, 2020.
Mechanical Science and Technology, vol. 36, no. 10, [21] Q. Meng, C. Chen, P. Wang, Z. Sun, and B. Li, “Study on
pp. 4913–4923, 2022. vehicle active suspension system control method based on
[6] K. H. Floreán-Aquino, M. Arias-Montiel, J. Linares-Flores, homogeneous domination approach,” Asian Journal of
J. G. Mendoza-Larios, and A. Cabrera-Amado, “Modern Control, vol. 23, no. 1, pp. 561–571, 2021.
semi-active control schemes for a suspension with MR ac- [22] S. Fazeli, A. Moarefanpur, and A. Moarefanpur, “An
tuator for vibration attenuation,” Actuators, vol. 10, no. 2, adaptive approach for vehicle suspension system control in
p. 22, 2021. presence of uncertainty and unknown actuator time delay,”
[7] J. Yang, D. Ning, S. Sun et al., “A semi-active suspension using Systems Science & Control Engineering, vol. 9, no. 1,
a magnetorheological damper with nonlinear negative- pp. 117–126, 2021.
8 Mathematical Problems in Engineering

[23] J. Zhao, X. Wang, P. K. Wong, Z. Xie, J. Jia, and W. Li, “Multi- [39] T. A. Nguyen, “Improving the comfort of the vehicle based on
objective frequency domain-constrained static output feed- using the active suspension system controlled by the double-
back control for delayed active suspension systems with integrated controller,” Shock and Vibration, vol. 2021, Article
wheelbase preview information,” Nonlinear Dynamics, ID 1426003, 11 pages, 2021.
vol. 103, no. 2, pp. 1757–1774, 2021. [40] M. L. Nguyen, T. T. H. Tran, T. A. Nguyen, D. N. Nguyen, and
[24] G. Wang, M. Chadli, and M. V. Basin, “Practical terminal N. D. Dang, “Application of MIMO control algorithm for
sliding mode control of nonlinear uncertain active suspension active suspension system: a new model with 5 state variables,”
systems with adaptive disturbance observer,” IEEE, vol. 26, Latin American Journal of Solids and Structures, vol. 19, no. 2,
no. 2, pp. 789–797, 2021. 2022.
[25] H. Chen, Y. J. Liu, L. Liu, S. Tong, and Z. Gao, “Anti-satu- [41] H. Pang, Y. Chen, J. Chen, and X. Liu, “Design of LQG
ration-based adaptive sliding-mode control for active sus- controller for active suspension without considering road
pension systems with time-varying vertical displacement and input signals,” Shock and Vibration, vol. 2017, Article ID
speed constraints,” IEEE Transactions on Cybernetics, vol. 52, 6573567, 13 pages, 2017.
no. 7, pp. 6244–6254, 2022. [42] Y. Guo, B. Wang, A. Tkachev, and N. Zhang, “An LQG
[26] B. Lin, X. Su, and X. Li, “Fuzzy sliding mode control for active controller based on real system identifcation for an active
suspension system with proportional diferential sliding mode hydraulically interconnected suspension,” Mathematical
observer,” Asian Journal of Control, vol. 21, no. 1, pp. 264–276, Problems in Engineering, vol. 2020, Article ID 6669283, 10
2019. pages, 2020.
[27] S. Kilicaslan, “Control of active suspension system consid- [43] Q. Zhu, J. J. Ding, and M. L. Yang, “LQG control based lateral
ering nonlinear actuator dynamics,” Nonlinear Dynamics, active secondary and primary suspensions of high-speed train
vol. 91, no. 2, pp. 1383–1394, 2018. for ride quality and hunting stability,” IET Control Teory &
[28] D. N. Nguyen, T. A. Nguyen, and N. D. Dang, “A novel sliding Applications, vol. 12, no. 10, pp. 1497–1504, 2018.
mode control algorithm for an active suspension system [44] D. Rodriguez-Guevara, A. Favela-Contreras, F. Beltran-Car-
considering with the hydraulic actuator,” Latin American bajal, D. Sotelo, and C. Sotelo, “Active suspension control
Journal of Solids and Structures, vol. 19, no. 1, 2022. using an MPC-LQR-LPV controller with attraction sets and
[29] R. Bai and D. Guo, “Sliding-mode control of the active quadratic stability conditions,” Mathematics, vol. 9, no. 20,
suspension system with the dynamics of a hydraulic actuator,” p. 2533, 2021.
[45] R. Bai and H. B. Wang, “Robust optimal control for the vehicle
Complexity, vol. 2018, Article ID 5907208, 6 pages, 2018.
suspension system with uncertainties,” IEEE Transactions on
[30] T. A. Nguyen, “Advance the efciency of an active suspension
Cybernetics, 2022.
system by the sliding mode control algorithm with fve state
[46] C. Wei, K. Zhang, Y. Cai, Z. Wang, and W. Yu, “A new
variables,” IEEE Access, vol. 9, Article ID 164368, 2021.
method of static output-feedback H∞ controller design for 5
[31] H. D. Choi and S. H. You, “Fuzzy fnite memory state esti-
DOF vehicle active suspension system,” Journal of the Bra-
mation for electro-hydraulic active suspension systems,” IEEE
zilian Society of Mechanical Sciences and Engineering, vol. 40,
Access, vol. 9, Article ID 99364, 2021.
no. 3, p. 132, 2018.
[32] J. Sun, J. Cong, W. Zhao, and Y. Zhang, “Quantized feedback
[47] M. G. Unguritu, T. C. Nichitelea, and D. Selisteanu, “Design
control of active suspension systems based on event trigger,”
and performance assessment of adaptive harmonic control for
Shock and Vibration, vol. 2021, Article ID 8886069, 15 pages,
a half-car active suspension system,” Complexity, Article ID
2021. 3190520, 14 pages, 2022.
[33] J. Liu, X. Li, X. Zhang, and X. Chen, “Modeling and simulation [48] W. Gang, “ESO-based terminal sliding mode control for
of energy-regenerative active suspension based on BP neural uncertain full-car active suspension systems,” International
network PID control,” Shock and Vibration, vol. 2019, Article Journal of Automotive Technology, vol. 21, no. 3, pp. 691–702,
ID 4609754, 8 pages, 2019. 2020.
[34] J. Luo, H. Pang, M. Wang, and R. Yao, “PI observer-based [49] G. I. Y. Mustafa, H. Wang, and Y. Tian, “Optimized fast
fault-tolerant tracking controller for automobile active sus- terminal sliding mode control for a half-car active suspension
pensions,” IEEE Access, vol. 10, Article ID 47203, 2022. systems,” International Journal of Automotive Technology,
[35] M. Haddar, R. Chaari, S. C. Baslamisli, F. Chaari, and vol. 21, no. 4, pp. 805–812, 2020.
M. Haddar, “Intelligent PD controller design for active sus- [50] A. R. Moghadam, H. Kebriaei, and H. Kebriaei, “Stochastic
pension system based on robust model-free control strategy,” sliding mode control of active vehicle suspension with mis-
Proceedings of the Institution of Mechanical Engineers - Part C: matched uncertainty and multiplicative perturbations,” Asian
Journal of Mechanical Engineering Science, vol. 233, no. 14, Journal of Control, vol. 22, no. 6, pp. 2330–2339, 2020.
pp. 4863–4880, 2019. [51] Y. J. Liu and H. Chen, “Adaptive sliding mode control for
[36] S. Y. Han, J. F. Dong, J. Zhou, and Y. H. Chen, “Adaptive fuzzy uncertain active suspension system with prescribed perfor-
PID control strategy for vehicle active suspension based on mance,” IEEE Transactions on Systems, Man, and Cybernetics:
road evaluation,” Electronics, vol. 11, no. 6, p. 921, 2022. Systems, vol. 51, no. 10, pp. 6414–6422, 2021.
[37] H. Metered, W. Abbas, and A. S. Emam, Optimized Pro- [52] F. Zhao, M. Dong, Y. Qin, L. Gu, and J. Guan, “Adaptive
portional Integral Derivative Controller of Vehicle Active neural-sliding mode control of active suspension system for
Suspension System Using Genetic Algorithm, SAE Technical camera stabilization,” Shock and Vibration, vol. 20158 pages,
Paper, Warrendale, PA, USA, 2018. Article ID 542364, 2015.
[38] Z. A. Karam and O. A. Awad, “Design of active fractional PID [53] A. Azizi and H. Mobki, “Applied mechatronics: designing a
controller based on whale’s optimization algorithm for sta- sliding mode controller for active suspension system,”
bilizing a quarter vehicle suspension system,” Periodica Pol- Complexity, vol. 2021, Article ID 6626842, 23 pages, 2021.
ytechnica Electrical Engineering and Computer Science, vol. 64, [54] D. N. Nguyen and T. A. Nguyen, “A novel hybrid control
no. 3, pp. 247–263, 2020. algorithm sliding mode-PID for the active suspension system
Mathematical Problems in Engineering 9

with state multivariable,” Complexity, vol. 2022, Article ID [70] X. Shao, J. Zhang, L. Xu, and W. Zhang, “Appointed-time
9527384, 14 pages, 2022. guaranteed adaptive fault-tolerant attitude tracking for
[55] H. Wang, L. Chang, and Y. Tian, “Extended state observ- quadrotors with aperiodic data updating,” Aerospace Science
er–based backstepping fast terminal sliding mode control for and Technology, vol. 132, 2022.
active suspension vibration,” Journal of Vibration and Con-
trol, vol. 27, no. 19-20, pp. 2303–2318, 2021.
[56] M. Jafari, S. Mobayen, F. Bayat, and H. Roth, “A nonsingular
terminal sliding algorithm for swing and stance control of a
prosthetic leg robot,” Applied Mathematical Modelling,
vol. 113, pp. 13–29, 2023.
[57] S. Mobayen, F. F. M. El-Sousy, K. A. Alattas, O. Mofd,
A. Fekih, and T. Rojsiraphisal, “Adaptive fast-reaching
nonsingular terminal sliding mode tracking control for
quadrotor UAVs subject to model uncertainties and external
disturbances,” Ain Shams Engineering Journal, Article ID
102059, 2022.
[58] V. Ghafari, S. Mobayen, S. ud Din, T. Rojsiraphisal, and
M. T. Vu, “Robust tracking composite nonlinear feedback
controller design for time-delay uncertain systems in the
presence of input saturation,” ISA Transactions, vol. 129,
pp. 88–99, 2022.
[59] C. Ozbek, O. C. Ozguney, R. Burkan, and N. Yagiz, “Design of
a fuzzy robust-adaptive control law for active suspension
systems,” Sadhana, vol. 45, no. 1, p. 194, 2020.
[60] H. Li, Z. Zhang, H. Yan, and X. Xie, “Adaptive event-
triggered fuzzy control for uncertain active suspension
systems,” IEEE Transactions on Cybernetics, vol. 49, no. 12,
pp. 4388–4397, 2019.
[61] B. Shaer, J. P. Kenne, C. Kaddissi, and C. Fallaha, “A chat-
tering-free fuzzy hybrid sliding mode control of an electro-
hydraulic active suspension,” Transactions of the Institute of
Measurement and Control, vol. 40, no. 1, pp. 222–238, 2018.
[62] Y. Zhang, Y. Liu, and L. Liu, “Adaptive fnite-time NN control
for 3-DOF active suspension systems with displacement
constraints,” IEEE Access, vol. 7, Article ID 13577, 2019.
[63] T. Wang and Y. Li, “Neural-network adaptive output-feed-
back saturation control for uncertain active suspension sys-
tems,” IEEE Transactions on Cybernetics, vol. 52, no. 3,
pp. 1881–1890, 2022.
[64] R. Saleh, “A comparative study of particle swarm optimized
control techniques for active suspension system,” International
Review of Automatic Control, vol. 15, no. 4, pp. 213–221, 2022.
[65] A. K. Abdul Zahra and T. Y. Abdalla, “Design of fuzzy super
twisting sliding mode control scheme for unknown full ve-
hicle active suspension systems using an artifcial bee colony
optimization algorithm,” Asian Journal of Control, vol. 23,
no. 4, pp. 1966–1981, 2021.
[66] K. Wu and C. Ren, “Control and stability analysis of double
time-delay active suspension based on particle swarm opti-
mization,” Shock and Vibration, vol. 2020, Article ID 8873701,
12 pages, 2020.
[67] X. Shao, J. Zhang, and W. Zhang, “Distributed cooperative
surrounding control for mobile robots with uncertainties and
aperiodic sampling,” IEEE Transactions on Intelligent
Transportation Systems, 2022.
[68] J. Zhang, X. Shao, W. Zhang, and J. Na, “Path-following
control capable of reinforcing transient performances for
networked mobile robots over a single curve,” IEEE Trans-
actions on Instrumentation and Measurement, 2022.
[69] W. Zhang, X. Shao, W. Zhang, J. Qi, and H. Li, “Unknown
input observer-based appointed-time funnel control for
quadrotors,” Aerospace Science and Technology, vol. 126,
Article ID 107351, 2022.